He et al. (2023) SMPD: a soil moisture-based precipitation downscaling method for high-resolution daily satellite precipitation estimation

Identification

• Journal: Hydrology and earth system sciences
• Year: 2023
• Authors: Kunlong He, Wei Zhao, Luca Brocca, Pere Quintana Seguí
• DOI: 10.5194/hess-27-169-2023

Research Groups

• Institute of Mountain Hazards and Environment, Chinese Academy of Sciences, Chengdu, China
• School of Civil Engineering, Sun Yat-sen University, Guangzhou, China
• Research Institute for Geo-Hydrological Protection, National Research Council, Perugia, Italy
• Ebro Observatory (OE), Ramon Llull University – CSIC, Roquetes, Spain

Short Summary

This study introduces the Soil Moisture-based Precipitation Downscaling (SMPD) method, a novel physically-based approach utilizing the soil-water balance equation and high-resolution soil moisture to downscale daily GPM IMERG precipitation from 10 kilometers to 1 kilometer over the Iberian Peninsula, resulting in enhanced spatial detail and improved accuracy metrics.

Objective

• To establish a Soil Moisture-based Precipitation Downscaling (SMPD) scheme that uses high-resolution surface soil moisture (SSM) and Normalized Difference Vegetation Index (NDVI) as auxiliary variables to generate fine-scale (1 kilometer) daily precipitation estimates from coarse-resolution satellite products (GPM IMERG), thereby overcoming the limitations of coarse spatial resolution in hydrological applications.

Study Configuration

• Spatial Scale: Downscaling from 10 kilometers (0.1°) to 1 kilometer (0.01°) resolution. The study area is the Iberian Peninsula (Southwestern Europe).
• Temporal Scale: Daily resolution. The study period spans 3 years, from 2016 to 2018.

Methodology and Data

• Models used: Soil Moisture-based Precipitation Downscaling (SMPD) method, which is based on the inversion and simplification of the soil-water balance equation (inspired by SM2RAIN). The method uses an adaptive window approach for local calibration of model coefficients and Kriging interpolation for residual correction to ensure value preservation.
• Data sources:
  • Satellite Precipitation: Integrated Multi-satellite Retrievals for GPM (IMERG) V06B Final Run daily precipitation product (0.1° spatial resolution).
  • Auxiliary Data: European Space Agency (ESA) Climate Change Initiative (CCI) Surface Soil Moisture (SSM) product (0.25°, downscaled to 1 kilometer); MODIS/Terra 16-day Normalized Difference Vegetation Index (NDVI) product (1 kilometer).
  • Validation Data: Daily precipitation measurements from 1027 rain gauge stations provided by the Spanish State Meteorological Agency (AEMET).

Main Results

• The downscaled 1 kilometer daily IMERG product showed improved performance compared to the original 10 kilometer product when validated against 1027 rain gauges.
• Daily Accuracy: The Correlation Coefficient (CC) increased slightly from 0.60 to 0.61. The Root Mean Square Error (RMSE) decreased from 4.99 millimeters to 4.83 millimeters, and the relative Bias (BIAS) decreased from 9% to 5%.
• Rainfall Event Detection: The downscaled results significantly improved event detection: Probability of Detection (POD) increased from 0.84 to 0.88; False Alarm Ratio (FAR) decreased from 0.52 to 0.47; and Critical Success Index (CSI) increased from 0.44 to 0.48.
• Spatial Detail: The SMPD method successfully enhanced spatial heterogeneity, providing more detailed precipitation distribution information and removing the blocky appearance of the coarse-scale pixels.
• Monthly Accuracy: When aggregated to the monthly scale, the downscaled data maintained high accuracy with a CC of 0.84, an RMSE of 30.88 millimeters, and a BIAS of 5%.

Contributions

• Physically-Based Downscaling: Developed a novel downscaling method (SMPD) based on the physical principle of the surface water balance equation, providing a more solid physical foundation compared to traditional empirical or statistical downscaling techniques.
• Use of High-Resolution SSM: This is one of the first downscaling methods to successfully incorporate high-resolution daily Surface Soil Moisture (SSM) as a key driving factor, leveraging its strong physical connection with precipitation.
• Independence and Robustness: The method is independent of rain gauge density for the downscaling process (gauges are only used for validation), making it highly applicable to regions with sparse meteorological networks.
• Daily Scale Improvement: Demonstrated high effectiveness and robustness in improving precipitation accuracy and spatial detail at the daily scale, a temporal resolution often challenging for physically-based downscaling methods.

Funding

• National Natural Science Foundation of China (grant nos. 42071349 and 42222109)
• Sichuan Science and Technology Program (Grant No. 2020JDJQ0003)
• West Light Foundation of the Chinese Academy of Sciences
• Project PRIMA PCI2020-112043 (funded by MCIN/AEI/10.13039/501100011033)

Citation

@article{He2023SMPD,
  author = {He, Kunlong and Zhao, Wei and Brocca, Luca and Quintana‐Seguí, Pere},
  title = {SMPD: a soil moisture-based precipitation downscaling method for high-resolution daily satellite precipitation estimation},
  journal = {Hydrology and earth system sciences},
  year = {2023},
  doi = {10.5194/hess-27-169-2023},
  url = {https://doi.org/10.5194/hess-27-169-2023}
}

Generated by BiblioAssistant using gemini-flash-latest (Google API)